The results of the analysis are summarised in Fig. 3. Overall, composting achieved the best score for 7 out of 14 environmental and health impacts, while AD had the highest environmental burden in 6 impacts and the mean ranking of the three scenarios (1 = best, 3 = worst) were compost: 1.7, AD: 2.1, and incineration: 2.3. A hotspot analysis that shows the contribution of different life cycle stages of each scenario is shown in Fig. 4.
For composting, the operation stage contributed the most to the majority of impacts, in particular depletion of fossil fuels (60%), terrestrial acidification (60%), terrestrial eutrophication (80%), and ozone depletion (90%). The environmental impacts of compost could be offset by its use as a substitute to synthetic fertiliser, the most significant being climate change (55%), freshwater eutrophication (95%), depletion of abiotic resources elements (48%). The treatment of food waste contributes to the majority of environmental burdens associated with AD. These impacts are attributed primarily to two processes, namely the use of auxiliary materials and wastewater treatment. Construction has a slight impact especially regarding eco-toxicity at 17% and carcinogenic human toxicity carcinogenic at 30%. Most burdens are offset by the use of recovered energy to substitute conventional energy sources (i.e. natural gas and coal).
Incineration has a similar environmental hotspot source to those of composting and AD: operation is the stage with the highest environmental burden while energy recovery (both heat and electricity) plays a key role in offsetting these burdens. The significant contribution from operation is the auxiliary materials used in order to control emissions, ammonia injection to control NOx emissions, lime for control of SO2 and HCL and activated carbon to capture heavy metals.
The supremacy of composting’s results in the majority of the investigated impacts is largely due to the nature of composting itself: it is a technologically simple process that requires considerably low energy and auxiliary material inputs compared to either AD or incineration. In addition, the analysis results show that the majority of the burden in AD and incineration is associated with the treatment of by-products, namely wastewater and methane purification for AD, and Air Pollution Control (APC) for incineration.
The exclusion of the impacts of food collection and transportation in this study also works in favour of composting and AD, as these options require a separate collection system of food waste and, therefore, additional energy input. When treated via incineration, food waste is co-collected and does not require an additional separate food waste collection system. Therefore, incineration consumes less energy input compared to composting and AD. Food collection and transportation are significant for the depletion of fossil fuels, elements, and the ozone layer (Jeswani and Azapagic 2016). A separate food collection system would increase the environmental burden across these categories (Burnley et al. 2011). The lack of food collection and transportation in this study is due to this study being a gate-to-grave assessment, meaning food waste collection and transportation were not included in the analysis. This was due to the fact that the goal of the assessment was to investigate the environmental impacts of food waste management technologies rather than that of the overall waste management system. Future studies could include this larger system boundary in their analysis.
Although the results of the analysis tend to show that composting is more environmentally friendly than other options, it performs worse than AD in two key environmental impacts: climate change and depletion of fossil fuels. This is attributed primarily to the fact that, unlike AD and incineration options, composting process does not include the generation of energy (in the form of electricity or heat) and, therefore, does not offset the huge quantities of emissions as AD and incineration do. The management of 1 tonne of food waste in a composting facility could emit 74 kg CO2-eq./FU while the AD and incineration options achieve significant benefits for the same functional unit; estimated average reductions in AD and incineration being − 2400 kg CO2-eq. and − 3000 kg CO2-eq., respectively. This unsurprising result is attributed to the fossil fuel energy substituted by the production of energy (in the form of electricity, heat, or both) in both AD and incineration options. If a decarbonised energy mix is substituted, the impacts of composting are reduced.
We have assumed the composition and water content of UK food waste via previous studies (Table A.7). As the UK decarbonises, the UK diet may change. This may affect the quality of UK food waste as a feedstock—specifically the water content. A variation in water content will have a large impact on the three treatment methods studied. In case of Incineration, there might not be any positive energy output if the moisture content of food waste goes beyond 50% since more energy is spent in drying which would directly affect the ADP-F (Fig. 3). Similarly the C, H, N, O, S ratio will also have their direct impacts on the environmental benefits and burdens studied here. Future studies could further modify UK food waste and energy generation to match dietary transition.
This study has also not included a comparative cost analysis of the different treatment methods to complement its environmental analysis. Future research could provide a comparative cost analysis of treatment methods under existing and potential decarbonisation scenarios. Furthermore, the study covers three well-established treatment technologies; other food waste treatment technologies are in their “infancy” in the UK waste market and therefore have been excluded. Future research could also broaden scope to compare other waste treatment methods under existing and potential decarbonisation scenarios.
Hybrid LCA vs truncation error
The results of the study demonstrate the ability of the hybrid LCA model to include additional processes and, therefore, reduce truncation error, compared to a conventional process-based LCA method. The adoption of the hybrid model increases estimates of associated GHG emissions with composting, AD and incineration by approximately 26, 10 and 11%, respectively. Thus, the analysis results quantitatively confirm the ability of the hybrid model to reduce truncation error by expanding boundaries of the modelled scenario (Finnveden et al. 2009; Jeswani et al. 2010).
Additional environmental burdens captured are mainly associated with managerial and indirect services such as services of head offices, consulting services, third-party technical services, and testing and maintenance services. These service-based activities are difficult to model using LCA (and, therefore, ignored in the literature), yet contribute substantially to the overall environmental impacts of modelled scenarios. The hybrid model uses the expenditure cost of these activities to estimate their environmental impacts. These expenditure costs are estimated to constitute up to 12% for AD, 10% for compost and 34% for incineration in the construction stage. Expenditure on these services increases for day-to-day operation activities; it accounts for 40% of total expenditure on AD, 67% for compost and 44% for incineration.
The environmental impacts of capital goods
Results show that capital goods constitute up to 20% of the overall environmental impacts in 10 categories for compost and 4 categories for both AD and INC. The study findings support those of previous studies highlighting the significant environmental impacts of capital goods (Finnveden et al. 2005; Brogaard and Christensen 2016). The overall environmental burdens associated with capital goods are presented in Figure A.5.
Capital goods used for the composting scenario contributed mostly to the potential impacts on carcinogenic human toxicity (70%), climate change (45%) and eco-toxicity (40%). These impacts were caused primarily due to the use of cement and steel: the energy input in the production stage of these products contribute the most to climate change while the disposal of steel and cement slag is responsible for the emission of heavy metals and, therefore, contributes to both human toxicity and eco-toxicity (Burchart-Korol 2013; Salas et al. 2016). Significant toxicity impacts are also reported in AD and incineration due to capital goods. The production of steel contributes to both human toxicity and eco-toxicity by 50 and 45% for AD and 13 and 10% for incineration, respectively. In addition, a substantial impact on depletion of abiotic resources (elements) is estimated for both AD and incineration scenarios. This could be explained by the huge quantities of capital goods (in particular steel) required in the construction stage of AD and incineration as these infrastructures are large-scale construction projects compared to composting (Brogaard and Christensen 2016). Significant results were also reported in freshwater eutrophication impacts, in particular in the incineration scenario. This highest impact across all 14 environmental and health categories is caused by excessive use of copper, stainless steel and cement. This finding agrees with those of Brogaard and Christensen (2016) that report a 90% freshwater eutrophication impact due to capital goods.
Comparison with previous literature
Figure 5 provides comparison of ten LCA studies (including this study), reporting greenhouse gas emissions per tonne of food waste processed. The previous studies reported a large variation in the GWP and some broad patterns emerge. Four out of six studies conclude that incineration with energy recovery has the greatest environmental benefits while only two studies conclude in favour of AD. This could be explained due to the high energy input required to operate an AD facility (compared to incineration) and the additional diesel consumption required to set up a separate food waste collection system (Burnley et al. 2011). In addition, variations in the biogas yield and the type of substituted energy adopted in reviewed studies contribute significantly to these discrepancies. For example, Eriksson et al. (2015) reports AD results which are 4 times larger than those reported in this study. This substantial difference could be attributed due to the assumption made by Eriksson and his colleagues that the entire theoretical yield of biogas was produced, while this study is based on actual AD plant figures (see Table 2) Eriksson et al.’s study also assumes that biogas replaces diesel as a fuel for city buses, while this study assumes biogas substitutes UK electricity production by natural gas (61.5%) and coal (38.5%).
Robustness of results
The sensitivity of results was investigated by way of a Monte Carlo Analysis. Overall results, listed in Table 3 below and plotted as error bars in Fig. 3, reveal a high level of uncertainty that could lead to a change in the ranking of scenarios in 7 impact categories between AD and incineration, 4 impact categories between AD and composting, 6 impacts between incineration and composting, and 2 impact categories amongst all scenarios studied (i.e., freshwater eutrophication and ozone depletion). The high level of uncertainty makes it difficult to draw a generic conclusion. However, this study has helped to better understand environmental impact patterns. This analysis also highlights the importance of the quality of data used in order to conduct an environmental assessment of a specific technology.
Notwithstanding the large variability in some parameters as shown in Table 3, the indicator values for all metrics are statistically significant from one another (p < 0.01), except for the effect of composting and anaerobic digestion on marine eutrophication and non-carcinogenic toxicity.
Results indicate that the decarbonisation of the UK supply electricity mix leads to reduction in overall environmental burdens across impact categories (Fig. 6). Scenario 4 (Ref100) has reduced the overall impact of global warming potential (GWP) by 82%, while scenario 2 (Ref 65-A) leads to 50% reduction. With regard to Scenario 3 (Ref 65-B), the substitution of natural gas by nuclear energy as a source of low-GHG energy leads to an additional 16% reduction in GWP but does increase the environmental burden in four impact categories: non-carcinogenic human toxicity, ionising radiation (IR), depletion of elements (ADP-E) and particulate matter.